A Predictive Approach for Selecting Suitable Computing Nodes in Grid Environment by Using Data Mining Technique - Advances in Computational Science and Engineering - page 204

Fig. 4. The ratio of successfully finished jobs in both Random method and our provider
6 Conclusion and Future Works
Dynamic and Instantaneous Selecting some nodes for scheduling in the pool of dis-
covered nodes is a challenging problem. Many methods for this purpose have been
presented but all of them have some restrictions. We have shown that our method is a
viable contender for use in future Grid implementations. This is supported by the
experimental results obtained on a Grid. Our proposed approach is learning based
with high accuracy which can reduced extra overhead communications and faults in
cycle of selection. This broker provider application along with MA offer a dynamic
decision to access any of the available and appropriate nodes by using main important
criteria. Considering this characteristic we recommend this method is useful in cases
in which we have sufficient nodes. For the future work we decide use a dynamic
framework for the broker on economic-based grid. In this way, depend on staus of
jobs and grid, The broker can select a specific nodes with considration of cost, time
execution,expiration time, and so on.
Acknowledgments. This work was funded by Islamic Azad University-Miyandoab
branch. The authors would like to thanks to Mr. Ali Imani and Mr. Hosein
Rajabalipoure at University Technology of Malaysia (UTM), for their helps in
preparation of this paper.
References
1. Czajkowski, K., Fitzgerald, S., Foster, I., Kesselman, C.: Grid information services for dis-
tributed resource sharing. In: 10th IEEE Symposium on High Performance Distributed
Computing, San Francisco, California, August 7-9 (2001)
2. Krauter, K., Buyya, R., Maheswaran, M.: A Taxonomy and Survey of Grid resource Man-
agement Systems. Software Practice and Experience 32(2), 135-164 (2002)